Glial fibrillary acidic protein (GFAP) is one of the major structural proteins in astrocytes, and its expression is markedly up-regulated following injury. To explore the role of GFAP in astrocytes, during the previous funding period the investigator generated knockout and transgenic mice that are either completely deficient in or express elevated levels of, GFAP. The GFAP-null mice display only subtle alterations of astrocyte morphology, and are relatively normal unless stressed by disease or trauma. In contrast, the GFAP over-expressing mice develop astrocytes that are enlarged and contain complex inclusion bodies identical to the Rosenthal fibers seen in Alexander's disease. Mice in high expressing lines died before weaning. The phenotype of GFAP over-expressing mice raises important questions about how Rosenthal fibers form and how they affect astrocyte function. In addition, these results lead to a specific prediction regarding a genetic basis for Alexander's disease. During the next funding period , the investigator proposes to further explore the significance of GFAP over-expression by the following set of experiments. In specific aim 1, he will identify the molecular requirements for Rosenthal fiber formation by crossing the GFAP transgene into mice that are unable to express the endogenous mouse GFAP, vimentin, or alphaB-crystallin (three of the major components of Rosenthal fibers). In addition, he will determine whether it is over-expression per se, or the expression of two distinct GFAP polypeptides (mouse and human), that is responsible for Rosenthal fibers. In specific aim 2, he will test the hypothesis that Alexander's disease in humans results from a primary abnormality in the GFAP gene, by analyzing genomic DNA from patients and controls for mutations and gene duplications. In specific aim 3, he will test the hypothesis that formation of Rosenthal fibers in astrocytes is associated with fundamental physiological changes in astrocytes and in neuron-glial interactions.